Abstract
We find all analytic surfaces in 3dimensional Euclidean space such that through each point of the surface one can draw two transversal circular arcs fully contained in the surface (and analytically depending on the point). The search for such surfaces traces back to the works of Darboux from XIXth century. We prove that such a surface is an image of a subset of one of the following sets under some composition of inversions:

the set \(\{p+q:p\in \alpha ,q\in \beta \}\), where \(\alpha ,\beta \) are two circles in \(\mathbb {R}^3\);

the set \(\{2\frac{p \times q^{}}{p+q^2}:p\in \alpha ,q\in \beta ,p+q\ne 0\}\), where \(\alpha ,\beta \) are circles in the unit sphere \({S}^2\);

the set \(\{(x,y,z): Q(x,y,z,x^2+y^2+z^2)=0\}\), where \(Q\in \mathbb {R}[x,y,z,t]\) has degree 2 or 1.
The proof uses a new factorization technique for quaternionic polynomials.
This is a preview of subscription content, log in to check access.
Notes
 1.
E.g., on issues with this very assertion in books by Hartshorne and Goerz–Wedhorn see: http://math.stackexchange.com/questions/70293/whydoesthedefinitionofanopensubschemeopenimmersionofschemesallowf, http://mathoverflow.net/questions/182044/whyistheinverseofabijectiverationalmapration.
References
 1.
Beauregard, R.: When is F[x, y] a unique factorization domain? Proc. Am. Math. Soc. 117(1), 67–70 (1993)
 2.
Beauville, A.: Complex algebraic surfaces, London Math. Soc. Student Texts, vol. 34, 2nd edn. Cambridge University Press, Cambridge, p. 132 (1996)
 3.
Brauner, H.: Die windschiefen Kegelschnittfläschen. Math. Ann. 183, 33–44 (1969)
 4.
Cass, D., Arpaia, P.J.: Matrix generation of Pythagorean ntuples. Proc. Am. Math. Soc. 109(1), 1–7 (1990)
 5.
Coolidge, J.L.: A Treatise on the Circle and Sphere, p. 603. The Clarendon Press, Oxford (1916)
 6.
Degen, W.: Die zweifachen Blutelschen Kegelschnittflächen. Manuscr. Math. 55(1), 9–38 (1986)
 7.
Dietz, R., Hoschek, J., Juettler, B.: An algebraic approach to curves and surfaces on the sphere and on other quadrics. Comput. Aided Geom. Design 10, 211–229 (1993)
 8.
Eilenberg, S., Niven, I.: The “fundamental theorem of algebra” for quaternions. Bull. Am. Math. Soc. 50, 246–248 (1944)
 9.
Gelfand, I., Gelfand, S., Retakh, V., Wilson, R.L.: Quasideterminants. Adv. Math. 193, 56–141 (2005)
 10.
Gentili, G., Stoppato, C., Struppa, D.: Regular Functions of a Quaternionic Variable. Springer Monographs in Math. Springer, Berlin, Heidelberg, p. 185 (2013)
 11.
Gentili, G., Stoppato, C.: Zeros of regular functions and polynomials of a quaternionic variable. Mich. Math. J. 56, 655–667 (2008)
 12.
Gohberg, I., Lancaster, P.J., Bodman, L.: Matrix Polynomials. Academic Press Inc, Cambridge (1982)
 13.
Guth, L., Zahl, J.: Algebraic curves, Rich Points, and DoublyRuled Surfaces. arXiv:1503.02173
 14.
Gordon, B., Motzkin, T.S.: On the zeros of polynomials over division rings. Trans. Amer. Math. Soc. 116, 218–226 (1965)
 15.
Ivey, T.: Surfaces with orthogonal families of circles. Proc. Am. Math. Soc. 123(3), 865–872 (1995)
 16.
Kocik, J.: Clifford algebras and Euclid’s parametrization of Pythagorean triples. Adv. Appl. Clifford Algebras 17(1), 71–93 (2007)
 17.
Kollár, J.: Rational Curves on Algebraic Varieties. Springer, Berlin, Heidelberg (1996)
 18.
Kollár, J.: Quadratic Solutions of Quadratic Forms. arXiv:1607.01276
 19.
Krasauskas, R., Zubė, S.: Rational Bézier formulas with quaternion and Clifford algebra weights. In: Dokken, T., Muntingh, G. (eds.) SAGA–Advances in Shapes, Geometry, and Algebra, Geometry and Computing, vol. 10, pp. 147–166. Springer, New York (2014)
 20.
Lavicka, R., O’Farrell, A.G., Short, I.: Reversible maps in the group of quaternionic Moebius transformations. Math. Proc. Camb. Philos. Soc. 143(1), 57–69 (2007)
 21.
Lubbes, N.: Minimal families of curves on surfaces. J. Symb. Comput. 65, 29–48 (2014). arXiv:1302.6687
 22.
Lubbes, N.: Families of circles on surfaces. arXiv:1302.6710
 23.
Lubbes, N.: Clifford and Euclidean Translations of Circles. arXiv:1306.1917v11
 24.
Nilov, F., Skopenkov, M.: A surface containing a line and a circle through each point is a quadric. Geom. Dedicata 163(1), 301–310 (2013). arXiv:1110.2338
 25.
Ore, O.: Theory of noncommutative polynomials. Ann. Math. (II) 34, 480–508 (1933)
 26.
Pakharev, A., Skopenkov, M.: Surfaces containing two circles through each point and decomposition of quaternionic matrices. Russ. Math. Surv. 72(2(434)), 195–196 (2017). arXiv:1510.06510
 27.
Postnikov, M.: The Fermat Theorem. An Introduction to Algebraic Numbers Theory. Nauka, Moscow, p. 128 (1978)
 28.
Pottmann, H., Shi, L., Skopenkov, M.: Darboux cyclides and webs from circles. Comput. Aided Geom. D. 29(1), 77–97 (2012). arXiv:1106.1354
 29.
Schicho, J.: The multiple conical surfaces. Contrib. Algebra Geom. 42(1), 71–87 (2001)
 30.
Shafarevich, I.R.: Algebraic Geometry, vol. 1. Springer, New York (1974)
 31.
Smertnig, D.: Factorizations of elements in noncommutative rings: a survey. In: Chapman, S., Fontana, M., Geroldinger, A., Olberding, B. (eds) Multiplicative Ideal Theory and Factorization Theory. Springer Proceedings in Mathematics and Statistics, vol. 170. Springer, Cham (2016)
 32.
Takeuchi, N.: Cyclides. Hokkaido Math. J. 29, 119–148 (2000)
 33.
Timorin, V.A.: Rectifiable pencils of conics. Mosc. Math. J. 7(3), 561–570 (2007)
Acknowledgements
This results have been presented at Moscow Mathematical Society seminar, SFB “Discretization in geometry and dynamics” colloquim in Berlin, Ya. Sinai—G. Margulis conference, and the conference “Perspectives in real geometry” in Luminy. The authors are especially grateful to A. Pakharev for joint numerical experiments, related studies [26], and for pointing out that numerous surfaces we tried to invent all have form (4). The authors are grateful to N. Lubbes and L. Shi for parts of Fig. 1, to A. Gaifullin and S. Ivanov for finding gaps in earlier versions of the proofs, to A. Bobenko, J. Capco, S. Galkin, A. Kanunnikov, O. Karpenkov, A. Klyachko, J. Kollár, W. Kühnel, A. Kuznetsov, N. Moshchevitin, S. Orevkov, F. Petrov, R. Pignatelli, F. Polizzi, H. Pottmann, G. Robinson, I. Sabitov, J. Schicho, K. Shramov, S. Tikhomirov, V. Timorin, M. Verbitsky, E. Vinberg, J. Zahl, S. Zubė for useful discussions. The first author is grateful to King Abdullah University of Science and Technology for hosting him during the start of the work over the paper.
Author information
Affiliations
Corresponding author
Additional information
Dedicated to the last real scientists, searching only for the truth, not career, not glory, not pushing forward their own field or students.
The article was prepared within the framework of the Academic Fund Program at the National Research University Higher School of Economics (HSE) in 2015–2016 (Grant No 15010092) and supported within the framework of a subsidy granted to the HSE by the Government of the Russian Federation for the implementation of the Global Competitiveness Program. The first author was also supported by the President of the Russian Federation Grant MK6137.2016.1, “Dynasty” foundation, and the Simons–IUM fellowship. The second author was partially supported by the MarieCurie Initial Training Network SAGA, FP7PEOPLE contract PITNGA214584.
Communicated by Thomas Schick.
Appendix. A rationality result
Appendix. A rationality result
In this appendix we give a direct proof of Lemma 4.8 going along the lines of [29] and exposed for nonspecialists. We use notation of Sect. 4, in particular, from the statements of lemmas there.
A projective algebraic surface \(\Psi \) is unirationally ruled (or simply uniruled), if there are an algebraic curve \(\eta \) and a rational map \(\eta \times P^1\dasharrow \Psi \) whose image is the whole \(\Psi \) possibly besides a 1dimensional set. A surface \(\Psi \) is birationally ruled (or simply ruled), if there is a birational map \(\eta \times P^1\dasharrow \Psi \). A curve \(\eta \) is rational, if there is a birational map \(P^1\dasharrow \eta \).
Lemma 5.15
The surface \({\hat{\Phi }}\) is unirationally ruled.
Proof
Let \({\hat{\gamma }}\) be the irreducible curve in the variety of conics given by Lemma 4.5. Consider the algebraic set \({\hat{\Psi }} := \{(P, \alpha )\in {\hat{\Phi }}\times {\hat{\gamma }} : P\in \alpha \}. \) The second projection \({\hat{\Psi }}\rightarrow {\hat{\gamma }}\) is a rational map such that a generic fiber is a conic (hence a rational curve). By the Noether–Enriques theorem [2, Theorem III.4] \({\hat{\Psi }}\) is birationally ruled. In particular there is an algebraic curve \(\eta \) and a rational map \(\eta \times P^1\dasharrow {\hat{\Psi }}\), whose image is the whole \({\hat{\Psi }}\) besides a 1dimensional set. Compose the map with the first projection \({\hat{\Psi }}\rightarrow {\hat{\Phi }}\). This projection is surjective by [30, Theorem 2 in §I.5.2] because the image contains the 2dimensional subset \(\bigcup _v \alpha _v\cap \Omega =\Phi \cap \Omega \) by Lemma 4.2. We get a rational map \(\eta \times P^1\dasharrow {\hat{\Phi }}\), whose image is \({\hat{\Phi }}\) besides a 1dimensional set, i.e., \({\hat{\Phi }}\) is unirationally ruled. \(\square \)
Lemma 5.16
Each unirationally ruled surface is birationally ruled.
Proof
Let \({\hat{\Phi }}\) be a unirationally ruled surface and \(\eta \times P^1\dasharrow {\hat{\Phi }}\) be a rational map whose image is the whole \({\hat{\Phi }}\) besides a 1dimensional set. Take a desingularization \(d:{\tilde{\Phi }}\rightarrow {\hat{\Phi }}\) defined in the 1st paragraph of the proof of Lemma 4.9. Let \({\hat{\Phi }}\dasharrow {\tilde{\Phi }}\) be the inverse rational map of the desingularization.
Consider the rational map \(\eta \times P^1\dasharrow {\hat{\Phi }}\dasharrow {\tilde{\Phi }}\). By the theorem on eliminating indeterminacy [2, Theorem II.7] this rational map equals a composition \(\eta \times P^1\dasharrow {\tilde{\Psi }}\rightarrow {\tilde{\Phi }}\), where \({\tilde{\Psi }}\) is a smooth projective algebraic surface, the first map is birational, and the second map is rational and defined everywhere. In particular, \({\tilde{\Psi }}\) is birationally ruled. Since the image \(\eta \times P^1\dasharrow {\hat{\Phi }}\) is the whole \({\hat{\Phi }}\) besides a 1dimensional set and the surfaces are compact, by [30, Theorem 2 in §I.5.2] it follows that \({\tilde{\Psi }}\rightarrow {\tilde{\Phi }}\) is surjective.
By the Enriques theorem [2, Theorem VI.17 and Proposition III.21], a smooth projective algebraic surface is birationally ruled if and only if for each \(k>0\) the kth tensor power of the exterior square of the cotangent bundle has no sections except identical zero (in other terminology, the surface has Kodaira dimension\(\infty \), or all plurigeni vanish). Assume, to the contrary, that \({\tilde{\Phi }}\) has such a section (pluricanonical section). Then the pullback under the surjective rational map \({\tilde{\Psi }}\rightarrow {\tilde{\Phi }}\) is such a section for the birationally ruled surface \({\tilde{\Psi }}\), a contradiction. Thus \({\tilde{\Phi }}\), and hence \({\hat{\Phi }}\), is birationally ruled. \(\square \)
Proof of Lemma 4.8
By Lemmas 5.15 and 5.16 the surface \({\hat{\Phi }}\) is birationally ruled. Thus there is a birational map \({\hat{\Phi }}\dasharrow \eta \times P^1\). Consider the two conics through a generic point of the surface \({\hat{\Phi }}\). Their images are two distinct rational curves through a point of \(\eta \times P^1\). Since there is only one \(P^1\)fiber through each point, at least one of the rational curves is nonconstantly projected to the curve \(\eta \). By the Lüroth theorem [2, Theorem V.4] the curve \(\eta \) must be rational, and hence \({\hat{\Phi }}\) is rational. \(\square \)
Remark 5.17
We conjecture that the following generalization of Lemma 4.8 is true: an algebraic surface containing two rational curves through almost every point is rational. See [17, Definition IV.3.2, Theorems IV.5.4, IV.3.10.3, IV.2.10, Corollary IV.5.2.1, Exercise IV.3.12.2] for a sketch of the proof. \(\square \)
Rights and permissions
About this article
Cite this article
Skopenkov, M., Krasauskas, R. Surfaces containing two circles through each point. Math. Ann. 373, 1299–1327 (2019). https://doi.org/10.1007/s002080181739z
Received:
Revised:
Published:
Issue Date:
Mathematics Subject Classification
 51B10
 13F15
 16H05